mixture models, monte carlo, bayesian updating and dynamic models mike west computing science and...
TRANSCRIPT
![Page 1: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/1.jpg)
Mixture Models, Monte Carlo, Bayesian Updating and
Dynamic Models
Mike WestComputing Science and
Statistics, Vol. 24, pp. 325-333, 1993
![Page 2: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/2.jpg)
Abstract
• The development of discrete mixture distributions as approximations to priors and posteriors in Bayesian analysis– Adaptive density estimation
![Page 3: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/3.jpg)
Adaptive mixture modeling• p() : the continuous posterior density function fo
r a continuous parameter vector .• g() : approximating density for importance sampl
ing function.– T-distribution
= {j, j=1,…,n} : random sample from g(). = {wj, j=1,…,n} : weights
– wj = p()/(kg())
– k = )(/)(1 j
n
j j gp
![Page 4: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/4.jpg)
Importance sampling and mixture
• Univariate random sampling– Direct Bayesian interpretations (based on mixt
ures of Dirichlet processes)• Multivariate kernel estimation
– Weighted kernel estimator
(1) ),|()( 2
11 hdwg j
n
jj Vθθθ
![Page 5: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/5.jpg)
Adaptive methods of posterior approximation
• Possible patterns of local dependence exhibited by p() – Easy
• Different regions of parameter space are associated with rather different patterns of dependence.– V is varying with local j and more heavily
depending on j.
![Page 6: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/6.jpg)
Adaptive importance sampling
• The importance sampling distribution is sequently revised based on information derived from successive Monte Carlo samples.
![Page 7: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/7.jpg)
AIS algorithm1. Choose an initial importance sampling
distribution with density g0(), draw a small sample n0 and compute weights, deducing the summary 0 = {g0, n0, 0, 0}. Compute the Monte Carlo estimates and V0 of the mean and variance of p0
2. Construct a revised importance function g1() using (1) with sample size n0, points 0,j, weights w0,j, and variance matrix V0
3. Draw a larger sample of size n1 from g1(), and replace 0 with 1
4. Either stop, and base inferences on 1, or proceed, if desired, to a further revised version g2(), constructed similarly.
0
![Page 8: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/8.jpg)
Approximating mixtures by mixtures
• The computational burden increases if further refinement with larger sample sizes.– Solution) Using a mixtures of several
thousand T
• Reducing the number of components by replacing ‘nearest neighboring’ components with some form of average
![Page 9: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/9.jpg)
Clustering routine1. Set r = n, starting with the r = n component mix
ture, choose k < n as the number of components for the final, reduced mixture.
2. Sort r values of j. in in order of increasing values of weights wj in
3. Find the index i such that j. is the nearest neighbor of 1, and reduce the sets and to sets of size r –1 by removing components 1 and i, and inserting ‘average’ values
i
iii
www
wwww
1*
111* )/()( θθθ
![Page 10: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/10.jpg)
4. Proceed to (2), stopping here only when r = k
5. The resulting mixture, the locations based on the final k averaged values, with associated combined weights, the same scale matrix V but new, and larger, window-width h based on the current, reduced ‘sample size’ r rather than n
![Page 11: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/11.jpg)
Sequential updating and dynamic models
• Updating a prior to posterior distribution for a random quantity or parameter vector based on received data summarized through a likelihood function for the parameter
![Page 12: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/12.jpg)
Dynamic models
• Observation model
• Evolution model
)|(~)|( 0 tttt YpY θθ
)|(~)|( 11 ttett p θθθθ
![Page 13: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/13.jpg)
Computations
• Evolution step– Compute the current prior for t.
• Updating step– Observing Yt, compute the current posterior
11111 )|()|()|( tttttett dDppDp θθθθθ
)|()|()|( 01 tttttt YpDpDp θθθ
![Page 14: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/14.jpg)
Computations: evolution step1. Various features of the prior p(t|Dt-1) of interest
can be computer directly using the Monte Carlo structure
2. The prior density function can be evaluated by Monte Carlo integration at any point
1
1,1,11 ]|[]|[
tn
iitteittt EwDE θθ
1
1,1,11 )|()|(
tn
iitteittt pwDp θθ
![Page 15: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/15.jpg)
3. The initial Monte Carlo samples t* (by t
from p(t| t-1,i)) provide starting values for the evaluation of the prior.
4. t* may be used with weights t-1 to
construct a generalized kernel density estimate of the prior
5. Monte Carlo computations can be performed to approximate forecast moments and probabilities
*
]|[]|[ 0,11
tt
ttittt YEwDYEθ
θ
![Page 16: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/16.jpg)
Computations: updating step
• Adaptive Monte Carlo density
![Page 17: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/17.jpg)
Examples
• Example 1– A normal, linear, first-order polynomial
model
• Example 2– Not normal– Using T distributions
• Example 3– bifurcating
]1),1(255.0[~)|(
]10,05.0[~)|(2
1111
2
ttttt
ttt
N
NY
![Page 18: Mixture Models, Monte Carlo, Bayesian Updating and Dynamic Models Mike West Computing Science and Statistics, Vol. 24, pp. 325-333, 1993](https://reader036.vdocuments.us/reader036/viewer/2022082711/56649ef65503460f94c09657/html5/thumbnails/18.jpg)
Examples
• Example 4– Television advertising